Research
Research
For a complete list and access to my coauthored publications in peer-reviewed journals click here.
Network Design
The design of efficient networks make our lives convenient and functional. Depending on the application, these networks may be subject to topological constraints (e.g. cannot contain a cycle) or must consider additional costs such as capacity loading and transportation costs.
Although there are several algorithms that quickly calculate network designs that are at most a known percentage away from peak performance, mathematical programming offers a way to find the network design with the guaranteed best performance. The limitation however, is that the mathematical programs for network design are very large and require a significant amount of time to find the best design.
To address this, our team has focused on exploiting problem specific structure and decomposition methods to significantly reduce the time required by state-of-the-art mathematical programming methods bringing us closer to solving instances of large-scale that appear in industrial applications of network design.
Related Papers
C.A. Zetina, I. Contreras, E. Fernández, C. Luna-Mota. Solving the Optimum Communication Spanning Tree Problem , European Journal of Operational Research.
C.A. Zetina, I. Contreras, J.F. Cordeau. Exact Algorithms for the Multicommodity Uncapacitated Fixed-charge Network Design Problem.
C.A. Zetina, I. Contreras, J.F. Cordeau. Profit-oriented network design with elastic demand.
Hub Location
Hub and spoke networks provide an efficient means to collect, consolidate and transport goods from several origins and destinations around the world. Airline, freight and mail delivery companies are only a few examples of those who have adopted this network structure as their means of operation with slight variations depending on the specific goods to be transported.
Given the large number of origins and destinations and potential hub facility locations, designing these networks by means of mathematical programming methods is often a time consuming task that rarely culminates in finding the best hub network configuration. In addition, different aspects of the resulting hub and spoke network must be considered and improved for it to be a viable means of transportation.
We investigate hub network characteristics desirable for certain industries and propose mathematical programming models and algorithms to obtain them in significantly shorter computing times than the state-of-the-art.
Related Papers
C.A. Zetina, I. Contreras, J.F. Cordeau, E. Nikbakhsh. Robust uncapacitated hub location , Transportation Research Part B: Methodological.
I. Contreras, S. Jayaswal, N. Vidyarthi, C.A. Zetina. An exact algorithm for single-source quadratic facility location. Working paper.